White-headed Woodpecker monitoring on the Barry Point Fire for the Lakeview Stewardship CFLR, Fremont-Winema National Forest, 2013 progress report Finalized February 2014 by: USFS Rocky Mountain Research Station Victoria Saab, Jonathan Dudley, and Quresh Latif To: Fremont-Winema National Forest Amy Markus, Forest Wildlife Biologist Introduction The Collaborative Forest Landscape Restoration (CFLR) program is a cooperative effort to increase the rate of restoration on our National Forests. Monitoring is a key component of the CFLR program and our work is designed to address how well CFLR projects are meeting their forest restoration and wildlife habitat conservation goals. The white-headed woodpecker (Picoides albolarvatus; WHWO) is a regional endemic species of the Inland Northwest and may be particularly vulnerable to environmental change because it occupies a limited distribution and has narrow habitat requirements in dry coniferous forests. Monitoring in CFLR projects, such as the Lakeview Stewardship CFLR project on the Fremont-Winema National Forest (FreWin), also contributes to other ongoing, regional efforts to monitor effectiveness of silvicultural and prescribed-fire treatments for white-headed woodpeckers throughout their range in Oregon, Idaho, and Washington. Vegetation and fuels data collection also support modeling of fire-climate impacts on future forest conditions and wildlife habitat suitability. To meet their various ecological needs, white-headed woodpeckers require heterogeneous landscapes characterized by a mosaic of open- and closed-canopied ponderosa pine forests (Wightman et al. 2010, Hollenbeck et al. 2011), which are expected to benefit vascular plant and other vertebrate wildlife populations (e.g., Noss et al. 2006). Consequently, monitoring white- headed woodpecker populations and their habitat associations is central to biological monitoring for the Lakeview Stewardship CFLR project on the FreWin National Forest. Prescribed burning and thinning treatments planned under this CFLR project are intended to improve the landscape heterogeneity required by WHWOs. Thus, the principal goal of monitoring is to verify the effectiveness of forest restoration treatments for improving habitat and populations of WHWO, and to evaluate the influence of postfire salvage logging. This monitoring effort contributes to answering two questions identified by the Lakeview CFLRP monitoring plan (Markus et al. in draft 2013): What are the site specific effects of restoration treatments on focal species’ habitat and populations within a project area? What are the effects of restoration treatments on focal species’ habitat across the CFLR Project Area? Approximately 8% of the CFLR project area burned in the 2012 Barry Point Fire, which burned 54,440 acres of National Forest lands in Oregon. We shifted our monitoring focus in 2013 to the Barry Point Fire because: (1) planned post-fire salvage logging potentially affects WHWO 1 populations within the CFLR; and (2) the Barry Point Fire provided an opportunity to field test and refine the WHWO habitat suitability model for burned forests (Wightman et al. 2010). Field testing and refinement of habitat suitability models are necessary to assess and verify their predictive value, and thereby establish their utility for guiding land management decisions and conservation planning (i.e., practicing adaptive management). This report describes the monitoring protocol, the data obtained during the first year, and future plans for monitoring. Objectives of the WHWO monitoring in the Barry Point Fire were to: (1) examine changes in WHWO occupancy before and after postfire salvage logging; (2) validate and refine habitat suitability models developed using nest location data from the 2002 Toolbox Fire (Wightman et al. 2010); (3) monitor the effectiveness of woodpecker reserves identified by applying Toolbox-Fire habitat suitability models; and (4) characterize habitat (trees and snags) at nest and non-nest random locations. Study Area & Methods The 2013 monitoring plan relied on estimating WHWO occupancy rates and surveying for nest locations in forests burned by the 2012 Barry Point Fire (Figure 1). We established five study units ranging from 164 - 439 ha based on predictions of habitat suitability models developed from the 2002 Toolbox Fire (Wightman et al. 2010; Appendix 1). Each study unit contained approximately equal proportions of predicted suitable and unsuitable WHWO nesting habitat, resulting in units containing a range of available nesting habitat within the Barry Point Fire. In addition, each study unit contained areas proposed for salvage logging following the 2013 breeding season, and areas designated as “woodpecker reserves”, where no logging would take place. The monitoring protocol for the Barry Point Fire deviated from the effectiveness monitoring protocol used for conventional CFLR treatments in unburned forest (Mellen-McLean et al. 2013) to meet multiple objectives stated above. The best approach to meet all objectives was to survey the established study units as completely as possible. Therefore, we conducted belt transect surveys that aimed for a complete survey of study units rather than point-based surveys described in the effectiveness monitoring protocol (Mellen-McLean et al. 2013). We determined WHWO occupancy from broadcast-call surveys and concurrent nest searches in each study unit. We conducted call-broadcast surveys along 200-m-wide belt transects covering the five study units (e.g., Figure 2) from 0600–1130 during the nesting season (visit 1: 25 April – 10 June; visit 2: 5 June – 9 July). We established a grid of guide points, which we used to guide transect surveys so that transects and ultimately study units were sampled completely and evenly. Guide points were spaced 189 m apart, aligned with transects, and distributed evenly across each study unit. During each survey, the surveyor zigzagged between guide points, stopping at each point to broadcast WHWO vocalizations aimed at eliciting responses by territorial woodpeckers. At each guide point, the surveyor played the broadcast for 2.5 minutes, listened for an additional 2-minute period of silence, and then moved on towards the next guide point until completing the transect. During subsequent visits to a given transect, surveyors followed an alternate set of guide points within that transect (Figure 2, inset). Surveyors recorded WHWO detections at guide points when conducting broadcast calls, as well as during the time when walking between guide points, so sampling effort was distributed continuously across belt transects. Surveyors estimated the location of the bird where first detected, then recorded the UTMs at that location, which could have been in surveyed or adjacent transects. Thus, using a priori established points as a guide, we 2 ensured the overall sampling effort was as complete, even, and redundant as possible to provide the highest quality data for fitting occupancy models. One belt-transect survey was typically completed in 1-3 mornings. On any given morning, surveyors were staggered across every other transect to avoid biasing woodpecker behavior. We recorded the date, time, UTM coordinates (NAD1983 zone 10N), and gender of each WHWO detected during transect surveys (see Appendix 2). Concurrent with standardized transect surveys, we conducted nest searches daily from April through July, meandering at minimum within belt-transect boundaries (i.e. 100 m to each side of the belt transect center; Dudley and Saab 2003), but often up to 1 km when following adults. Once a nest was located, we collected UTM coordinates and visited nests only until nest initiation was determined (i.e. presence of eggs or nestlings). Due to a limited budget, we did not determine nest fate as suggested for effectiveness monitoring (Mellen-McLean et al. 2013). We recorded nest locations of Black-backed Woodpeckers (Picoides arcticus; BBWO) that were found incidentally to our WHWO surveys, but did not actively search for BBWO. Following nest searches and broadcast-call surveys, we measured trees, snags, and stumps at 20 points within each study unit selected randomly from the guide points used for call-broadcast surveys and at all WHWO nest locations following a modification of the Region 6 (R6) WHWO monitoring protocol (Mellen-McLean et al. 2013; Appendix 3). We modified the protocol by increasing the plot size to 50 m x 20 m for all measurements to account for the rarity of live trees after wildfire. Despite substantial increases in snags subsequent to the fire, we maintained the R6 standard protocol plot size for snags to capture future changes in snag densities following potential salvage logging or restoration treatments. Data Analysis We will use occupancy models to examine changes in WHWO occupancy related to potential management activities. Occupancy models estimate species occurrence while correcting for our imperfect ability to detect the species of interest (MacKenzie et al. 2006). We will use 200×200 m cells super-imposed upon transect-survey data as the sampling unit for these models (Figure 2). Thus, models will estimate occupancy and detection probabilities for these cells. Occupancy models require the analyst to characterize sampling effort, allowing estimation of detection probabilities. We will use post-hoc analysis of WHWO-to-surveyor distances at the time of detection to extrapolate the area sampled. Surveying occupancy in both pre- and post-treatment units and in treated and untreated areas